Phylogeography of the subterranean rodent Ctenomys torquatus: an evaluation of the riverine barrier hypothesis

Abstract

Aim Subterranean rodents of the genus Ctenomys are widespread in open habitats; the collared tuco‐tuco, C. torquatus, has a wide range crossed by large water‐courses. Based on a phylogeographical approach to investigate the riverine barrier hypothesis, we evaluated the strength of a river as a barrier and characterized the effect of large rivers in structuring distinct populations of the collared tuco‐tuco. Location Grasslands of southern Brazil. Methods Sampling included 294 individuals from 33 localities throughout the geographical range of C. torquatus, especially around the main rivers. Phylogeographical patterns were estimated through two mitochondrial DNA sequences (control region and COI). Additionally, 22 microsatellite loci were surveyed to estimate the effect of rivers on population divergence through the FST index of genetic differentiation and AMOVAs. Spatial autocorrelation analyses were performed for both molecular markers. Molecular rates of change for mitochondrial DNA were estimated using fossil records and applied to Bayesian demographic analysis to test the relationship with the river geological data. Results Molecular‐clock analysis estimated the time to most recent common ancestor (TMRCA) to be c. 200 kyr and indicated a pattern of recent demographic expansion for C. torquatus. Fu's FS neutrality test was significant and negative (−13.7). High and significant FST and AMOVA among‐population comparisons (FSC) revealed highly differentiated populations, although no AMOVA groupings considering rivers as a segregating factor were significant. Spatial autocorrelation analysis revealed that isolation by distance was detected over approximately 200 km. Main conclusions The effect of the river was not proportionally larger from the headwaters to the mouth, and no population differentiation was observed after the later events of dispersal over the river or the formation of a water channel. This finding could result from the recent history of the occupation of C. torquatus in that region, combined with a dry Pleistocene climate and lower river levels.

Spatial and temporal divergence of the torquatus species group of the subterranean rodent Ctenomys

“…Within the torquatus group, phylogenetic relationships were recovered in concordance to previous studies (Freitas et al, 2012;Roratto et al, 2015;Caraballo et al, 2016). Namely, the (Corrientes group, C. pearsoni complex), (C. torquatus, (C. minutus, C. lami)), and C. ibicuiensis clades split apart from a basal polytomy ( Figure 2).…”

“…With low posterior support (0.66) C. sociabilis results sister to C. tuconax and together with the frater group are basal to the rest of the genus. These intergroup relationships had also been recovered by Roratto et al (2015).…”

“…Divergence times of the major lineages that comprise the speciose genus Ctenomys, have been estimated by Parada and collaborators (2011) on the basis of cyt-b sequences. However, the torquatus group was not completely represented in this study and, with the exception of C. torquatus, whose age has been recently estimated (Roratto et al, 2015), the origin of the main lineages of this highly diverse group has not been dated.…”

“…Regarding Ctenomys, we included 31 sequences representing the torquatus group and 58 additional sequences comprising more than 50 species of the genus (included in Parada et al, 2011 and references therein). We included representative samples from the torquatus group: one representative of each of the seven main clades of the Corrientes group (Caraballo et al, 2012), 8 sequences from different populations of the C. pearsoni complex (Tomasco and Lessa, 2007;Caraballo et al, 2016), 7 sequences representing the widespread distribution of C. torquatus (Roratto et al, 2015), 4 sequences from each of four localities of C. ibicuiensis (Freitas et al, 2012) and 4 sequences from C. minutus and the only cyt-b sequence available in the Genebank from C. lami (Lopes et al, 2013). The complete multi-calibrated dataset has 132 terminals.…”

“…Chromosomal divergence within each lineage could be biased by the number of sampled individuals and/or populations. However, this is not the case, since, for instance, C. torquatus has the highest number of studied populations (34) (Roratto et al, 2015) although its mKD value is relatively low (4). In the other extreme, 16 neighbouring C. lami demes harbor 26 different karyomorphs (Freitas, 2007) (Figure 4).…”